Modern software control systems must cope with uncertainty, manage dynamic environments, as well as provide greater flexibility. Distributed agent-based control systems are the ideal candidates to manage such dynamic and unpredictable application domains, like transportation systems. This paper shows how to implement a decentralized decision strategy, based on an extended version of the Probability Collective (PC) framework. Categories and Subject Descriptors I.2.11 [Artificial Intelligence]: Distributed Artificial Intelligence--Coherence and coordination, intelligent agents, multiagent systems General Terms Algorithms, Design, Experimentation Keywords Agents, coordination, distributed optimization, vehicle routing problem